Speech recognition systems convert spoken audio data into human-readable text. Speech recognition may be used in many applications, such as, but not limited to, close-captioning of multimedia, transcription of dictation, and conversion of voicemail messages to text for e-mail or instant messaging. In addition to merely recognizing raw words from speech, speech recognition systems may apply transformational techniques to convert the raw spoken word stream to a form more appropriate for display. For example, some words are capitalized, punctuation is added, or spacing and abbreviation rules are applied, etc. Different applications using speech recognition may require different formatting conventions.